Analysis And Research On Emotional Tendency Of Comment Texts Oriented To Peer Mutual Evaluation

In the Internet and information age,the amount of online data has grown exponentially,and the use of online education platforms and social media has greatly increased the depth and breadth of educational data.The emergence of large-scale educational datasets has created a specialized branch in the field of data mining—Educational data mining.A large part of the educational data exists in the format of unstructured text,mainly including text comments in the learning process,which contain valuable insights and attitudes of learners.fully utilized.Students’ peer evaluation is conducive to teachers’ better diagnostic feedback,not only can effectively reflect teachers’ teaching process and students’ own performance,but also play a significant role in improving learners’ learning effectiveness,learning motivation,self-efficacy,etc.Can have a positive impact on learners’ learning.However,there is no detailed conclusion on how the affective tendencies in peer peer assessment affect the learning process and how it affects the learning process.At the same time,analyzing a large number of texts is a huge and very difficult workload for teachers.By combing the current research situation at home and abroad,this study applies text sentiment analysis as an educational data analysis method,and combines Bandura’s ternary interaction theory to determine sentiment from three aspects: mutual evaluation environment,mutual evaluation individual and mutual evaluation behavior.Based on the analysis elements,a semester-long peer review was organized based on the course "3D Animation Application Technology" to construct a review text dataset.Then,using the method of word frequency and weight calculation of feature words,calculation of sentence sentiment value,and judgment of comment depth,the three aspects of emotional tendency factors are determined,and the influence relationship with the learning results is analyzed.The research results show that,in terms of the mutual evaluation environment,it is concluded that the focus and themes of peer peer evaluation are normative,creative,technical ability,aesthetics,and details.The details of the work,the visual effects and color matching of the work,and the model design of the work are the key aspects to judge the quality of the course work;in the mutual evaluation of individuals,there is a significant relationship between different emotional tendencies and academic performance,especially negative There is a significant relationship between emotion and positive emotion and excellent grades and qualified grades.There is also a significant relationship between different degrees of emotional disposition and academic performance.There is a significant relationship between high emotional disposition and excellent grades and qualified grades.In terms of mutual evaluation behavior,there are significant differences in the academic performance of students with different comment depths in the comments,there is a significant relationship between the deep evaluation and the excellent grade,and there is a significant relationship between the shallow evaluation and the qualified grade.,indicating that the deeper the evaluation depth,the better the academic performance,and the shallower the evaluation degree,the more general the academic performance.Finally,this study puts forward suggestions from the content and types of evaluation texts,the construction and collection of evaluation data sets,the organization of classroom teaching,the organization of the mutual evaluation process,and the evaluation rubric.